Molecular weight and subunit structure of bovine adrenal glucose-6

BOVINE ADRENAL GLUCOSE-6-PHOSPHATE DEHYDROGENASE. Diwan Singh , Phil G. Squire. International Journal of Peptide and Protein Research 1975 ...
0 downloads 0 Views 2MB Size
SUBUNITS OF BOVINE

GLUCOSE-6-P

DEHYDROGENASE

Molecular Weight and Subunit Structure of Bovine Adrenal Glucose-6-phosphate Dehydrogenaset Diwan Sin&$ and Phil G. Squire*

The bovine adrenal glucose-6-phosphate dehydrogenase used in these studies was highly purified and crystallized three times. The mean value and standard deviation of six measurements of the sedimentation coefficient within the concentration range of 2-6 mg/ml was 9.45 f 0.065 S. The diffusion coefficient obtained from boundary spreading experiments in the synthetic boundary cell was 3.30 X lo-’ cm2 sec-1. The partial specific volume calculated from a comparison of low-speed sedimentation equilibrium experiments in HzO and DzO was 0.737. This value is in good agreement with the value 0.736 calculated from the amino acid composition. Several methods of dissociating the enzyme into its subunits were tested, but only reduction and carboxymethylation in ABSTRACT:

S

ince the discovery of glucose-dphosphate dehydrogenase gl glucose-6-phosphate :NADP-oxidoreductase, EC 1.1.1.49) by Warburg and Christian (1931a,b), the enzyme has been isolated from brewer’s yeast (Kornberg, 1950; Noltmann et a/., 1961), human erythrocytes (Kirkman, 1962; Chung and Langdon, 1963), mammary glands of cows (Julian et al., 1961) and of guinea pigs (Andrews, 1962), Escherichia coli (Sanwal, 1970), rat liver (Matsuda and Yugari, 1967), bovine ovary (Nielson and Warren, 1965), adrenal glands (McKerns, 1964; Squire and Sykes, 1970), and many other biological systems. As the first enzyme of the hexose monophosphate pathway, this enzyme plays a key role in maintaining the level of the reduced coenzyme NADPH in the cell by catalyzing the following reaction Glc-6-P

+ NADP’ e

6-phosphogluconolactone

+ NADPH + H+

Control of this enzyme is of special importance in the bovine adrenal cortex since NADPH provided by glucose oxidation through the hexose monophosphate shunt is required for the synthesis of steroid hormones. Control by a variety of metabolites has in fact been reported (McKerns, 1963, 1964; Criss and McKerns, 1968b; McKerns and Kaleita, 1960; Schachet and Squire, 1971). Preliminary estimates of the molecular weight based on the sedimentation coefficient have been reported by Criss and McKerns (1968a), and by Squire and Sykes (1970), but no systematic determination of the parameters required for the precise calculation of

t From the Department of Biochemistry, Colorado State University, Fort Collins, Colorado 80521. Receiued August 17, 1973. This work was supported by National Science Foundation Grant GB-21425. I t has been assigned Scientific Series Number 1872 by the Colorado State University Experiment Station. I t was taken in part from a dissertation submitted by Mr. Singh to the Graduate School of Colorado State University for the degree of Doctor of Philosophy. Part of this work was presented at the 164th National Meeting of the American Chemical Society in New York City, Aug 1972, Abstract BIOL 204. $ Present address: Department of Biochemistry, University of Iowa, Iowa City, Iowa 52242.

0.1 sodium dodecyl sulfate produced essentially complete dissociation. The following parameters were obtained for the subunit: S ~ O =, ~3.70 S and D20,w= 5.33 X lo-’ cm2 sec-1. The molecular weight calculated from these values is 65,200. The value calculated from sedimentation equilibrium measurement is 63,900, and from polyacrylamide gel electrophoresis measurements in sodium dodecyl sulfate is 64,500. The best estimate of the molecular weight of the subunit therefore is 64,600. We conclude from these and other data that the enzyme is composed of four subunits that are indistinguishable by sedimentation and electrophoretic criteria, and that each subunit has a molecular weight of 64,600, and consists of a single polypeptide chain.

the molecular weight of this polysubunit enzyme has been reported. The initial studies conducted in this laboratory were directed toward the determination of the molecular weight of the native enzyme by sedimentation measurements. Analysis of these experiments showed that even though the dependence of the sedimentation coefficient and the weight average molecular weight on concentration was small, plots of log c us. X2from the sedimentation equilibrium measurements were nonlinear, and the deviation from linearity increased with protein concentration, We concluded that the enzyme at this protein concentration, 2-6 mg/ml, was reversibly associating, and that the predominant association constant was small, but not negligible. These experiments provided valuable information about the state of aggregation of the native enzyme, but did not permit characterization of the mass of the enzyme with the precision we wanted. The alternative was to completely dissociate the enzyme, and determine the number and molecular weight of the subunits under conditions where association equilibria were absent or negligible. Materials and Methods Nicotinamide adenine dinucleotide diphosphate (NADP+), maleic anhydride, and sodium dodecyl sulfate were obtained from Sigma Chemical Co. Urea and DzO (99.8Z) were purchased from Mallinckrodt Chemical Works and Stohler Isotope Chemicals, respectively. Eastman Organic Chemicals was the source of the chemicals used in polyacrylamide gel electrophoresis. Gdn.HCI1 Ultra Pure grade, used in some of our studies, was purchased from Mann Research Laboratories Inc. Fast Green FCF and dithioerythritol were the products of Fischer Scientific Co. and P-L Biochemicals Inc.. respectively. Protein standards used for the determination of the molecular weight of bovine adrenal glucose-6-phosphate dehydrogenase by sodium dodecyl sulfate gel electrophoresis were : 1

Abbreviation used is: Gdn .HCl, guanidine hydrochloride.

B I O C H E M I S T R Y , VOL.

13,

NO.

9, 1 9 7 4

1819

SINGH AND SQUIRE

bovine serum albumin from Armour Pharmaceutical Co., lot no. F 71703; ovalbumin, pepsin, and hemoglobin from Pentex Inc. ; y-globulin from Mann Research Laboratories Inc. Dimethyl suberimidate used in the cross-linking experiment was synthesized from 1,6-dicyanohexane according to the procedure described by Davies and Stark (1970). Urea was recrystallized from 70% ethanol by the method of Steinhardt (1938). All other chemicals were of reagent grade and used without further purification. Glucose-6-phosphate dehydrogenase was isolated from bovine adrenal glands according to the procedure of Squire and Sykes (1970). The enzyme was crystallized three times before using it in our studies. The concentrations of samples used in sedimentation studies were determined refractometrically by means of the synthetic boundary cell. The activity of the enzyme was determined as previously described by Squire and Sykes (1970). Reduction and carboxymethylation of Glc-6-P-dehydrogenase were conducted essentially as described by Crestfield et al. (1963), except dithioerythritol was substituted for 2mercaptoethanol. The alkylated protein was examined for free sulfhydryl groups with a Beckman amino acid analyzer. When necessary, a second cycle of reduction and carboxymethylation was carried out to assure complete alkylation of the protein. The reduced and carboxymethylated enzyme was stored at -30". In preparing samples for sodium dodecyl sulfate gel electrophoresis, the reduction and carboxymethylation were carried out in the presence of 0.1 % sodium dodecyl sulfate rather than 8 M urea as recommended by Crestfield et al. (1963). Dissociation of the enzyme with maleic anhydride was carried out essentially as described by Cohen and Rosemeyer (1969). Prior to treatment with maleic anhydride, the enzyme was dialyzed for at least 24 hr against 0.1 M potassium phosphate buffer, containing M EDTA and 0.15 M KCl at pH 8.0, to remove ammonium ions remaining from the crystallization step. After dialysis a small volume of 2.0 M dibasic potassium phosphate was added to the protein solution in order to neutalize the maleic acid produced during the course of the reaction. The protein solution was then pipetted onto solid maleic anhydride, amounting to 30 equiv of the lysine residues present. The reaction mixture was incubated at 37" with constant stirring of the solution. No enzyme activity was observed after incubation for 30 min. The solution was then prepared for sedimentation and electrophoresis studies by dialysis against 0.1 M potassium phosphate (pH 7.0), M EDTA, 0.15 M KCl, and 0.1 2-mercaptocontaining ethanol. The procedure used for the determination of the molecular weight of the subunit, as well as the reduced and carboxymethylated Glc-6-P-dehydrogenase, combines features of the methods of Weber and Osborn (1969) and of Hedrick and Smith (1968). Gel concentration varied from 5 to IO%, but the Bis : acrylamide ratio was maintained at 1 :30. Operating procedures and relative mobility calculations were done according to Weber and Osborn (1969), but molecular weights were calculated from the slope of log of relative mobility cs. acrylamide concentration according to Hedrick and Smith (1968). Gels were stained with 0.1% solution of Fast Green FCF dissolved in 10% acetic acid. After destaining they were scanned at 620 nm by the Gilford uv recorder, Model 2000, connected to a Beckman D U recording spectrophotometer. The procedure of Davies and Stark (1970) was used for cross-linking the subunits with dimethyl suberimidate. The

1820

BIOCHEMISTRY,

VOL.

13,

NO.

9, 1 9 7 4

best results as judged by the appearance of a maximum number of bands were obtained with an enzyme to reagent ratio of 1 :2.5. The enzyme treated with the bifunctional reagent was subjected to sodium dodecyl sulfate gel electrophoresis (7.5 gel) as described earlier. All samples analyzed by sedimentation analysis were first dialyzed for at least 24 hr against an appropriate buffer, and the dialysate was used as the reference solvent. All transfers of the solution and the dialysate were performed with care to prevent differential evaporation. The sedimentation velocity and the equilibrium experiments of the native and dissociated glucose-6-phosphate dehydrogenase were carried out at 20" in a Beckman Spinco Model E analytical ultracentrifuge equipped with a schlieren diaphragm, a Raleigh interference optical system, and a rotor-temperature control (RTIC) unit. A double-sector cell was used in all experiments. Fluorochemical FC-43 was used to provide a visible lower meniscus in the sedimentation equilibrium experiments. All measurements on the photographic plate were made with a Nikon Model 6C optical micro comparator equipped with a rotational stage. The optical constants of the centrifuge were checked with the solution of a known refractive index by a fringe count as well as with a calibration cell furnished by the manufacturer. A capillary-type synthetic boundary cell was used for the determination of the initial protein concentration, CO, and plots of fringe positions us. time were used to determine the hinge point, i.e., the radial position at which the concentration is independent of time during the approach to equilibrium. Sedimentation coefficients were determined and ~ by Schachman (1957). corrected to ~ ~ 0as, described The diffusion coefficient, D, of the enzyme was determined by two different methods. One approach was the measurement of boundary spreading in the double-sector synthetic boundary cell by means of the Rayleigh interference optical system. The diffusion coefficient was calculated according to the procedure outlined by Schachman (1957) from the slope of the plot of (AH/Ah)*(4Dt/Fz) with respect to the corresponding time, t. Both sequential- and symmetrical-fringe pairing methods were utilized in order to determine the value of the diffusion coefficient. Unpublished experiments we have performed by this method on standard proteins have yielded results agreeing within a few tenths of a percent with published values. The second method was based on measurements taken during the approach to the sedimentation equilibrium, and utilized the equation derived by Van Holde and Baldwin (1958)

Here, t = (Aceq - Act)/Aceq represents the departure of the system from sedimentation equilibrium. Act is the concentration change between the inner and outer limits of the solution, ra and rb, at the time, t. Aceq refers to the system at equilibrium a

= (co/Ac,,)

=

RT/M(l - Op)dP(rb - ra)

(2)

where co is the initial protein concentration determined by an independent experiment with the synthetic boundary cell, and U ( a ) = (1

+

1/482a2)

(3)

The apparent molecular weight of the protein, M , was calculated from the slope of the plots of log c with respect to r2 by

M=

(2,303)(2RT)(d log c/dr*) (1 - F p ) d

(4)

SUBUNITS OF BOVINE

TABLE I :

GLUCOSE-6-P

DEHYDROGENASE

Physical Properties of Bovine Adrenal Glucosed-phosphate Dehydrogenase in the Native State. A. The sedimentation coefficient Concn (mg/ml) Buffera 4.37 5.50 0 . 3 M (NH&S04 0.15 M KC1 2-5.0 6.00 0.10 M NaCl S Z O ,= ~ 9.45 =!= 0.065 S B. The diffusion coefficient ( D ~ o , ~ ) By boundary spreading (concn 3.57 mg/ml> C. Partial specific volume, 0 Sedimentation equilibrium in D 2 0 and HzO buffers From amino acid composition D. The molecular weight Sedimentation equilibrium 283)OOO Sedimentation and diffusion 285, OOO Best value 284,000

x

1013 S 9.44 9.44 9.46b 9.42

3.30 x

Characterization of the Native Enzyme. The partial specific volume, B, was determined by the method of Edelstein and Schachman (1967). Two sedimentation equilibrium experiments were carried out with temperature, angular velocity, and the protein concentration kept constant. The density, p, was varied by substituting DzOin one of the experiments for H2O. The partial specific volume was calculated by applying the following relationship

- [(d In c/dr2)D20/(dIn c/dr2)~,01 (5) P D ~ O - PH20[(dIn C/dr2)D20/(dIn C/dr2)~,01 k

where k is a measure of exchangeability of hydrogen atoms of the protein with deuterium of D20. Experiments to determine the relative weight increase due to deuterium exchange have been carried out with a number of proteins, and gave an average value of k equal to 1.0155 (Hvidt and Nielsen, 1966). The partial specific volume of the native bovine adrenal Glc-6P-dehydrogenase, calculated from the slopes of log c us. X * plots (Figure l), was found to be 0.737 ml/g. Preparations of freshly purified Glc-6-P-dehydrogenase were characterized by sedimentation velocity experiments performed at 20°, and at a speed of 59780 rpm. The buffer used in four of these experiments was 0.10 M sodium acetate (pH 5.50), containing 2 X 10-j M NADP+, 10-2 M 2-mercaptoethanol, M EDTA, and 0.15 M KCI. This buffer was chosen because it approximates the conditions of maximal stability of the native enzyme. The results obtained from sedimentation velocity experiments in other buffers also are shown in Table I. Since no concentration dependence was observed in the concentration range we studied, the mean ~ ) taken as value of 9.45 A 0.065 S of six experiments ( s ~ , , , was the best value of this parameter. Measurements of boundary spreading in a double-sector synthetic boundary cell gave a value of 3.30 x 10-7 cm2 sec-1 for the diffusion coefficient of this enzyme. The molecular weight calculated from s and D by the Svedberg equation was 285,000. This is in good agreement with the average value of 283,000 calculated from the sedimentation equilibrium experiments for the native complex (eq 4). The

cm2 sec-1

0.737 ml/g 0.736 ml/g

a Buffer used in these studies was 0.1 M sodium acetate (pH 5 . 9 , containing 2 2-mercaptoethanol with additions as indicated. Average of four experiments.

Results

~ C

x

M

NADP+,

M

EDTA, and

M

plots of log c us. X2 for the native enzyme were not strictly linear. Since the linearity improved as the concentration decreased, the non-linearity was attributed to self-association. It was clear that it would be difficult to determine a precise value of molecular weight by this approach ; nevertheless, an approximate value proved to be useful in determining the state of aggregation of the native enzyme. The molecular weight determined from the linear region of the plot corresponding to the lower values of concentration was 283,000. Dissociation of the Enzyme into Subunits and Its Physical Properties. Attempts were made to dissociate the enzyme by treatment with maleic anhydride, with sodium dodecyl sulfate, by lowering the pH to 1.3 in 0.20 M HCI-KCI buffer, and by reduction followed by carboxymethylation in 8 M urea, in 8 M Gdn.HCI, and in 0.1% sodium dodecyl sulfate. All of these treatments resulted in substantial, but incomplete, dis-

1.20-

-

1.10-

-

M H , O D,O

1.00-

(3

c

-

a 0.90-

LL

216

220

225 230 X' (cm2)

235

240

1: Determination of 8 from plots of logarithm of the concentration in fringes, c, us. X * , at sedimentation equilibrium. Here X = Fr, where F is the magnification factor and r the radial distance. The experiments were performed in HzOand D20 solutions containing 0.10 M sodium acetate (pH 5.50), 0.30 M ammonium M NADP+. Sedimentation equilibrium measuresulfate, and ments were performed on the native enzyme at 4908 rpm and 20". The protein concentration was 2.381 mg/ml. FIGURE

BIOCHEMISTRY,

VOL.

13,

NO.

9, 1 9 7 4

1821

SINGH AND SQUIRE

TABLE II :Physical Properties of Bovine Adrenal

Glucose-6-phosphate Dehydrogenase Subunit.

A. The sedimentation coefficient Treatment Maleic anhydride Reduced and alkylated in sodium dodecyl sulfate HC1-KC1 (PH 1.3) Best valueb B. The diffusion coefficient (D*o,wx lo7 cm2 sec-1) Boundary spreading Approach to sedimentation equilibrium C. The molecular weight Sedimentation equilibrium Sedimentation and diffusion Acrylamide gel electrophoresis D. Other parameters Frictional coefficient (f) Frictional ratio ( f / f o ) Stokes’ radius, Re a

&o,w

(SI

(Sn/sd)’”

3.67 3.73 3. 57a 3.70

4.13 4.03 4.08 Average =

5.30 MI

5.33

Average = 64,600

65,200 64,500 7.48 x lo-* g sec-1 1.48 39.5 x 10-8cm

Not corrected for charge effects. * Values obtained in HC1-KC1 buffer were not included in the best value for szo,w.

sociation of the native enzyme. While the polydispersity of these preparations precluded accurate determination of the molecular weight of the subunit by sedimentation equilibrium, reliable measurements of the sedimentation coefficients could be made. These are recorded in Table 11. The values of 3.73 and 3.67 S for the enzyme dissociated by reduction and carboxymethylation, and by maleylation, respectively, are in good agreement. The low value of 3.57 S obtained in the HC1KC1 buffer at pH 1.30 may be attributed to charge effects. The ratios of ( S , / S , ~ ~ recorded ’~ in Table I1 are all approximately equal to four. This provides strong evidence that the native enzyme is a tetramer. Essentially complete dissociation of the enzyme was achieved when the reduced and carboxymethylated sample was dialyzed against 0.10 M sodium phosphate buffer (pH 7.0), containing 0.1 Z sodium dodecyl sulfate, and incubated at 37”. The plot of log c us. X2, unlike that obtained for the native enzyme, was linear throughout the radial distance. In a more sensitive test for linearity, plots of fringe deviations from the regression line showed only random scatter. The linearity of the plot provides evidence for the monodisperse nature of the subunit and the purity of original sample as well. The sedimentation equilibrium experiment yields a value for buoyant molecular weight, M(l up), for the sodium dodecyl sulfate-protein complex. The extent of sodium dodecyl sulfate binding, expressed as grams of sodium dodecyl sulfate per gram of protein, was determined from the difference in refractive index between the undiluted dialyzed solution and the equilibrium dialysate as described by Olins and Warner (1967). McBain et af. (1939) report that micelle formation is negligible at the low concentrations of sodium dodecyl sulfate (O.lz,3.47 mM) used in this study. The synthetic boundary cell experiment gives a fringe number, Jo, proportional to the refractive index contribution of the sodium dodecyl sulfate-protein complex. The fraction of these fringes due to the enzyme was calculated from its known concentration, its specific refractive increment, (dn/dc)E,,, and the optical constants of the ultracentrifuge. This fraction is proportional to the contribution of the protein to the refractive index increment, (A~)E,,. From this value, the extent of sodium dodecyl sulfate binding, & (grams of sodium dodecyl sulfate (SDS) per gram of protein), was determined by the following relationship

1822

BIOCHEMISTRY V O, L . 1 3 ,

NO.

9, 1 9 7 4

82

=

(h)sDs(dn/dC)Enx (h)En,(dn/dC)sDs

(6)

where (dn/dc)Ene,the specificrefractiveincrement of the protein, was taken as 1.86 X ml/mg; (dn/dc)sDs, the specific refractive increment of sodium dodecyl sulfate, is 1.209 x ml/mg (Hersh and Schachman, 1958), and ( A n ) s ~ s , the refractive index increment due to the sodium dodecyl sulfate bound to the protein, was obtained by subtracting (A~)E,, from the ( h ) c o m p l e x . The amount of sodium dodecyl sulfate bound per gram of Glc-6-P-dehydrogenase was found to be 0.423 g. With the use of this value of &, and O2 = 0.863 ml/g (Anacker et af., 1964), the molecular weight of Glc-6-Pdehydrogenase in the complex, M p , was computed from the following expression derived by Hersh and Schachman (1958)

where the numerator is the buoyant molecular weight term calculated experimentally. Here Gp and Oz are the partial specific volumes of protein and sodium dodecyl sulfate, respectively. The molecular weight of the subunit calculated in this manner (eq 7) was also corrected for the amount of iodoacetamide covalently linked to the cysteine residues during carboxymethylation of the enzyme. The calculated value of the molecular weight of the subunit, after these corrections were made, was 63,900 (Table 11). The data used in calculating the diffusion coefficient from boundary spreading in a double sector synthetic boundary cell are plotted in Figure 2. The diffusion coefficient ( D z ~ , ~ ) calculated from the slope was 5.35 x lo-’ cm2 sec-’’. Figure 3 illustrates the results of the measurements taken during the approach to sedimentation equilibrium. The diffusion coefficient, 5.30 x 10-7 cm2 sec-1, estimated by this method, is in good agreement with the value obtained from boundary spreading (Table 11). The linearity of these plots for the determination of the diffusion coefficient (Figures 2 and 3) provides an additional evidence of homogeneity of the sample. The molecular weight of the subunit was also estimated by polyacrylamide gel electrophoresis in 0.1 sodium dodecyl sulfate. The denatured protein moved as a single band, again indicating that these subunits have similar or identical molec-

SUBUNITS OF BOVINE

GLUCOSE-6-P

DEHYDROGENASE

211)

I

I

I

I

I

I

1

20

an

60

80

too

iw

tun

w1

I

0

I

I

I

I

I

I

I

1

1 0 2 0 0 0 4 0 5 0 6 0 7 0 8 0 9 0

TIME (mln) FIGURE 2: Determination of the diffusion coefficient of the dissmiated enzyme from the measurement of boundary spreading in a doublelector synthetic boundary cell by " I S of a Raleigh interference optical system. The slope of this maDh vields the value of the diffusion coefficient.

ular weights. The log of relative mobility was plotted against the corresponding gel concentrations as recommended by Hediick and Smith (1968), and a linear relationship was obtained for all the proteins tested. The slopes of these lines were plotted as a function of their molecular weights and by interpolation the molecular weight of bovine adrenal Glc-6P-dehydrogenase was found to be 64,500 (Table II), in excellent agreement with the values determined by the sedimentation studies. Since the subunits had a tendency to aggregate below pH 6.0, higher pH buffers (pH 6.3, 7.3, and 8.6) were used in order to examine the electrophoretic homogeneity of the subunits. In all cases only one band was observed suggesting that all the subunits have essentially the same mobility in this range of pH. Davies and Stark (1970) have shown how dimethyl suberimidate mav be used to determine the number and the tvoe .. of subunits contained in an oligoprotein. If the native bovine adrenal glcuose-6-P-dehydrogenaseconsists of four identical subunits, four bands should be obtained on sodium dodecyl sulfate gel electrophoresis with the dimethyl suberimidate treated enzyme. As shown in Figure 4, this was the result obtained.

-

IR-

0

c.) 1.3-

9

1.2 1.1

LO

-

0.0

FIGURE

l o a t (mkr) 3: Determination of the diffusion coefficient of the enzyme

subunit from the measurements taken during the approach to the sedimentation equilibrium. The diffusion coefficient of the subunits was calculated from the slope acwrding to eq 1, derivd by Van Holde and Baldwin (1958). and Sykes, 1970), and 239,000 (Criss and McKerns, 1968a), were subject to the same uncertainty, hut in addition, these values were calculated from sedimentation cwfficients alone by means of approximate equations involving assumptions which, while useful, cannot be expected to yield results of high accuracy. Because of the complications mentioned above, we have chosen to determine the molecular weight of the native enzyme complex by accurately determining the molecular weight and number of subunits under conditions where dissociation is

(-1 TOP

ABSORBANCE (620 nm)

Discussion The sedimentation coefficient of the native enzyme under conditions approximating those of the adrenal, but measured at a protein concentration in the range of 2-6 mg/ml, is 9.45 + 0.065 S.This value is in good agreement with the value 9.47 S previously reported from this laboratory (Squire and Sykes, 1970), but is somewhat lower than the value 9.80 S reported by Criss and McKerns (1968a) for the enzyme from the same source. Measurement of the diffusion coefficient of the native enzyme by the boundary-spreading technique yielded the value, Dso,v = 3.30 x lo-' cm' sec-1. The molecular weight calculated from szo,rvand D20,rais 285,000. This is in good agreement with the value 283,000 obtained from measure. ments at sedimentation equilibrium. Since we had evidence of mild association equilibria under the conditions of these experiments, we were uncertain as to the significance of these values. The other values in the literature, M = 190,000(Squire

FIGURE 4: The cross-linking of the bovine adrenal glucose-6-phosphate dehydrogenase. Electrophoresis was performed in 7.5% polyacrylamide gel containing sodium dodecyl sulfate as described under the Methods section. Here MI,M,, Ma, and M. represent monomer, dimer, trimer, and tetramer species, respectively.

BIOCHEMISTRY,

VOL.

13,

NO.

9, 1 9 7 4

1823

SINGH AND SQUIRE

complete. We accomplished this by complete reduction and carboxymethylation in 8 M urea followed by dialysis against a buffer containing 0.1% sodium dodecyl sulfate at room temperature for 24 hr. Attempts to dissociate the enzyme by other agents and even our early attempts by this method were discouraging in that dissociation, while substantial, appeared to be incomplete, and the resulting preparation was polydisperse. We surmise that the denatured protein aggregates during removal of the urea by dialysis, and that subsequent complete dissociation requires incubation in 0.1 sodium dodecyl sulfate at room temperature for several hours. The subunit molecular weight was determined by sedimentation equilibrium, and by measurements of the sedimentation and diffusion coefficients. Our best value, 64,600, is the mean of the values obtained by these methods (Table 11). The good agreement of this value with the value 64,500 obtained by acrylamide gel electrophoresis is remarkable. All these data lead to the conclusion that the native enzyme consists of four subunits. The ratio of the molecular weight of the native enzyme to that of the subunit is 284,000/64,600 = 4.4. The ratio of the sedimentation coefficients raised to the 3/2 power is 4.08. (For a discussion of this formulation, see Squire et a/., 1968.) Acrylamide gel electrophoresis of the partially cross-linked enzyme gives four bands as expected for a tetrameric structure. The fact that the ratio of molecular weights exceeds 4.0 may be attributed to a tetramer-octamer equilibrium which is significant in this concentration range. We, therefore, take as the best value for the molecular weight of the native tetramer four times the molecular weight of the monomer, or 258,000. All the data concerning the subunits are consistent with the hypothesis that the four subunits are identical. The linearity of the plot of log c us. X2 of data at sedimentation equilibrium of the dissociated subunits rules out the possibility of two populations of subunits having substantially different molecular weights. Analysis of the photographs of the sedimentation velocity studies on the same preparation leads to the same conclusion. Acrylamide gel electrophoresis of the subunit at pH values 6.3, 7.3, and 8.6, after treatment with maleic anhydride, showed only one band indicating that all subunit derivatives had the same mobility at these p H values. Observations of sodium dodecyl sulfate gel eyperiments following cross-linking with dimethyl suberimidate reveal only four bands (Figure 4), and the mobilities are proportional to the log of the molecular weights in the ratio 1 :2 :3:4. These observations also indicate that the subunits are identical, and certainly rule out large differences in the charge to mass ratio. Experiments designed to reveal possible minor differences in the subunits are in progress. The application of conventional criteria of purity to reversibly associating systems is often a difficult task. In such a system, direct evidence for the contention that the macromolecular constituent consists of a single component in various degrees of association depending on the overall protein concentration would be complete dissociative to a single monomeric component. The preceding discussion describes this accomplishment although we recognize that sodium dodecyl sulfate binding could have obliterated possible intrinsic charge differences in the subunits. We would add in this regard that the peptide map obtained from a tryptic digest of this preparation is also consistent with the amino acid analysis based on a single monomeric compound of mol wt 64,500. The fact that subunits were prepared by complete reduction and carboxymethylation of the enzyme suggests that the sub-

1824

BIOCHEMISTRY,

VOL.

13,

NO.

9, 1 9 7 4

units consist of a single polypeptide chain. Even though our attempts to dissociate the enzyme with maleic anhydride, urea, Gdn. HCI, and low pH did not result in complete dissociation, well over 80% of the protein had a sedimentation coefficient in fair agreement with our best value, 3.70 S. These observations lend additional support to this contention. The translational frictional coefficient, A and the frictional ratio, ( f / f o ) ,were calculated by means of the equations RT

J'=,,=-

M(l - u,p) Ns

and fo = 6~q(3Mii,/4rrN)'/~

(9)

These are true molecular constants calculated from experimental data, and are free of questionable assumptions. Their interpretation in terms of shape and hydration, however, would either require additional experimental data from other physicochemical methods, or would require the use of assumed values for shape or hydration. This problem is discussed in detail by several authors. For a recent review, see Squire (1974). The frictional ratio, fiJ0, calculated from eq 8 and 9 is 1.48 for the monomer-sodium dodecyl sulfate complex. While this frictional ratio could correspond to a rather wide range of models, the assignment of reasonable values to some of the undetermined parameters permits calculation of molecular dimensions that are also reasonable and consistent with other physicochemical data on protein-sodium dodecyl sulfate complexes. Thus if we assume a hydration value of 0.2 g of HzO/g of protein (Edsall, 1953) and utilize our experimentally determined value of 0.423 g of bound sodium dodecyl sulfate per gTf protein, we calculate the values a = 73.6 A and b = 16.4 A for the major and minor axes of the equivalent prolate ellipsoid of revolution. In a systematic study of the viscosity of solutions of proteinsodium dodecyl sulfate complexes, Reynolds and Tanford (1970) concluded that the proteins studied all appeared to have minor axes ranging from 13.4 to 15.1 A. Our value is 16.4 A, a somewhat higher value. In view of the fact that we have used a different hydrodynamic method to characterize the protein-sodium dodecyl sulfate complex, and have studied a protein not included in the work by Reynolds and Tanford (1970), we consider the agreement with the conclusion of these authors as satisfactory, and favor the prolate model of axial ratio 4.5 as best representing the Glc-6-P-dehydrogenasesodium dodecyl sulfate complex. Acknowledgments The success of this project has depended upon highly purified bovine adrenal Glc-6-P-dehydrogenase which was prepared by Mrs. Harriet Sykes. We also thank Drs. J. R. Bamburg for sharing his expertise in polyacrylamide gel electrophoresis technique, and D. W. Hamar, for the use of his laboratory facilities. References Anacker, E. W., Rush, R. M., and Johnson, J. S. (1964), J . Phys. Chem. 68,81. Andrews, P. (1962), Nature (London) 196,36. Chung, A. E., and Langdon, R. G. (1963), J . Bioi. Chem. 238,2309. Cohen, P., and Rosemeyer, M. A. (1969), Eur. J. Biochem. 8,8.

SE IN OLUTATHIONE PEROXIDASE

Crestfield, A. M., Moore, S., and Stein, W. H. (1963), J. Biol. Chem. 238,622. Criss, W. E., and McKerns, K. W. (1968a),Biochemistry 7,125. Criss, W. E., and McKerns, K. W. (1968b), Biochemistry 7, 2364. Davies, G. E., and Stark, G. R. (1970), Proc. Nut. Acad. Sci. U.S. 66,651. Edelstein, S . J., and Schachman, H. K. (1967), J. Biol. Chem. 242,306. Edsall, J. T. (1953), Proteins I, 549. Hedrick, J. L., and Smith, A. J. (1968), Arch. Biochem. Biophys. 126,155. Hersh, R . T., and Schachman, H. K. (1958), Virology 6,234. Hvidt, A., and Nielsen, S. 0. (1966), Advan. Protein Chem. 21,287. Julian, G. R., Wolfe, R. G., and Reithel, F. J. (1961), J. Biol. Chem. 236,754. Kirkman, H. N. (1962), J. Biol. Chem. 237,2364. Kornberg, A. (1950), J. Biol. Chem. 182,805. Matsuda, T., and Yugari, Y. (1967), J . Biochem. (Tokyo) 61, 535. McBain, M. E., Dye, W. B., and Johnston, S. A. (1939), J. Amer. Chem. SOC.61,3210. McKerns, K. W. (1963), Biochim. Biophys. Acta 71,710. McKerns, K. W. (1964), Biochim. Biophys. Acta 90,357.

McKerns, K. W., and Kaleita, E. (1960), Biochem. Biophys. Res. Commun.2,344. Nielson, M. H., and Warren, J. C. (1965), Biochim. Biophys. Acta 97,532. Noltmann, E. A., Gubler, C . J., and Kuby, S. A. (1961), J. Biol. Chem. 236,1225. Olins, A. L., and Warner, R. C. (1967), J. Biol. Chem. 242, 4994. Reynolds, J. A., and Tanford, C. (1970), J. Biol. Chem. 245, 5161. Sanwal, B, D. (1970),J. Biol. Chem. 245,1626. Schachet, M. S., and Squire, P. G. (1971), Biochim. Biophys. Acta 227,491. Schachman, H. K. (1957), Methods Enzymol. 4,32. Squire, P. G. (1974), in Molecular Electrooptics, O’Konski, C. T., Ed., New York, N. Y., Marcel Dekker (in press). Squire, P. G., Moser, P., and O’Konski, C. T. (1968), Biochemistry 7,4261. Squire, P. G., and Sykes, H. B. (1970), Znt.J . Protein Res. 2,173. Steinhardt, J. (1938),J . Biol. Chem. 123,543. Van Holde, K. E., and Baldwin, R. L. (1958), J. Phys. Chem. 62,734. Warburg, O., and Christian, W. (1931a), Biochem. 2.238,131. Warburg, O., and Christian, W. (1931b), Biochem. 2.238,206. Weber, K., and Osborn, M. (1969),J . Biol. Chem. 244,4406.

Selenium as a Component of Glutathione Peroxidase Isolated from Ovine Erythrocytest Sang-Hwan Oh, Howard E. Ganther, and William G. Hoekstra*

ABSTRACT :

By a nine-step procedure, glutathione peroxidase (glutathione :H2O2oxidoreductase, EC 1.1 1.1.9) was purified 4000-fold from ovine erythrocytes to homogeneity as determined by polyacrylamide gel electrophoresis and gel filtration chromatography. During purification the selenium concentration per unit of protein increased 3000-fold indicating that this enzyme accounted for about 75 % of the total ovine erythrocyte selenium. The purified enzyme contained 0.34%selenium and the approximate molecular weight as determined by gel

P

revious studies in our laboratory (Rotruck et al., 1971) demonstrated that dietary selenite exerted a marked protective effect against oxidant damage to the membrane and hemoglobin of rat erythrocytes incubated in vitro in the presence of ascorbate or H2O2.This effect depended upon the presence of glucose in the incubation medium and was clearly distinguishable from the protective effects of vitamin E. The Seglucose relationship was shown to involve the need for glucose metabolism in the continued regeneration of reduced gluta-

t From the Department of Biochemistry and the Department of Nutritional Sciences, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, Wisconsin 53706. Receiued December 3, 1973. Research supported by the College of Agricultural and Life Sciences, University of Wisconsin-Madison, and by U . S. Public Health Service Program Grant No. AM-14881.

filtration and polyacrylamide gel electrophoresis was 88,000, giving a stoichiometry of 3.8 g-atoms of selenium/mol of glutathione peroxidase. This is considered to represent 4 gatoms of selenium/mol of enzyme, presumably one selenium per enzyme subunit since sodium dodecyl sulfate gel electrophoresis suggested the presence of 4 subunits/mol of enzyme. Glutathione peroxidase is the first positively identified selenoenzyme of animal tissues.

thione (GSH) in the erythrocyte and a role for Se in the utilization of this GSH in protecting the cell against oxidant damage (Rotruck et al., 1971, 1972). Because GSH protects the cell, at least in large part, by serving as the hydrogen donor in the glutathione peroxidase catalyzed reduction of H202 and various hydroperoxides (Mills and Randall, 1958 ; Clive and Obrien, 1968), we attempted to determine if Se is a necessary component of this enzyme. It was found that most of the 76Se of rat erythrocytes labeled in uivo by injecting rats 2-4 weeks earlier with Na76SeOsfollowed glutathione peroxidase activity through two highly effective purification steps, but we did not purify the rat enzyme to homogeneity (Rotruck et a[., 1973). In order to obtain a larger amount of glutathione peroxidase than was easily provided by rat blood, we selected ovine blood as a source material in our further attempts at B I O C H E M I S T R Y , VOL.

13,

NO.

9, 1 9 7 4

1825